A Grey Markov Prediction Method Optimized by GSA for Forecasting Agricultural Cold Chain Transportation Volume in China
Authors:
Preprint
- 15 October 2026
- preprint
- Published by Springer Nature in Research Square
Abstract
Accurate prediction of frozen agricultural products logistics transportation volume is crucial for optimizing supply chain resource allocation. To solve problems of small sample sizes, nonlinear trends, and multiple influencing factors in the field, this study proposes a hybrid IGSA-Markov-FAGM(1,N) as prediction method. The model integrates Fractional-Order Accumulation Generation Operator (F-AGO) into GM(1,N) framework to emphasize the priority of new information and enhances nonlinear fitting ability, and Markov chain is also introduced to correct residuals, thus enhancing accuracy effectively. Meanwhile, Improved Gravitational Search Algorithm (IGSA) is used for optimization of parameter setting such as fractional order, weight factor and Markov state division thresholds. To verify the theory, 2013–2020 data of five Chinese agricultural-related indicators (including Primary Industry Added Value, Tertiary Industry Employees, etc.) are used as the training set to predict the frozen agricultural products logistics transportation volume from 2021 to 2024. Compared with the other four prediction methods that lack precision improvement measures, proposed hybrid prediction strategy demonstrates advantages in terms of MAPE, MAE, and RMSE (MAPE of the hybrid strategy is 10.23%, while MAPE of the two methods using fractional-order prediction is 13.56% and 17.81%, and the MAPE of the ordinary GM method, regardless of whether Markov correction is applied or not, exceeds 50%). This study provides a reliable prediction tool for agricultural cold chain logistics management and offers scientific support for industrial policy formulation.Keywords
All Related Versions
This publication has 38 references indexed in Scilit:
- Prediction of homogenized few-group constants for pressurized water reactor assembly using a Deep Neural Network with Gaussian Process Residual CorrectionAnnals of Nuclear Energy, 2026
- Forecasting and spatial distribution analysis of cold chain logistics demand for fresh agricultural products in ChinaEnvironment, Development and Sustainability, 2025
- Prediction of compression modulus of Nantong fine-grained soil based on an interpretability-driven GSA-optimized CatBoost modelAdvanced Engineering Informatics, 2025
- HyperVIX: A GWO‐Optimized ARIMA‐LSTM Hybrid Model for CBOE Volatility Index (VIX) ForecastingJournal of Forecasting, 2025
- Mixed-frequency fractional derivative grey model for forecasting electricity consumptionApplied Mathematical Modelling, 2025
- Smart Forecasting of Carbon Prices Using Machine Learning and Neural Networks: When ARIMA Meets XGBoost and LSTMJournal of Forecasting, 2025
- A novel fractional grey forecasting model for high-tech industry growth trendsApplied Mathematical Modelling, 2025
- Experimental and numerical gust identification using deep learning modelsApplied Mathematical Modelling, 2024
- GSA for machine learning problems: A comprehensive overviewApplied Mathematical Modelling, 2021
- A prediction method using the grey model GMC(1,n) combined with the grey relational analysis: a case study on Internet access population forecastApplied Mathematics and Computation, 2005